iLab Neuromorphic Vision C++ Toolkit Downloads

The iLab Neuromorphic Vision C++ Toolkit is distributed freely under GPL in an effort to broaden the development of a comprehensive set of C++ classes to ease the implementation of computational neuroscience models for real-time applications.

The source tree is maintained using the Subversion revision control system. The main development platform is Linux but the core functionality (excluding interfacing with cameras, robots, etc) also works under Windows (using cygwin) and MacOS X.

See the documentation page for many publicly-available resources which will assist you in getting started.

Publicly available downloads:

• README
• COMPILING
• Programmer notes (also contains info about special features)
• Online doxygen documentation
• Special Linux packages required to compile some of our code
• Test image databases and also see crcns.org for our public database of human eye movements.
• Complete Mandriva Linux distribution with pre-compiled and fully-installed saliency software to run on your Windows, Mac, or Linux machine using VirtualBox.
• Great Matlab Saliency Toolbox by Dirk Walther at Caltech
• Maemo port to run on the small Nokia N810 internet tablet, check out this screenshot.
• Neovision2 annotated video dataset which you can use to train and test your models.
• Source Code SVN Repository, updated in real-time
• Daily tarball snapshot (updated every night): saliency-current.tar.bz2

Online documentation note: The HTML Online Documentation is your most useful source of information. Be sure, in particular, that you check the Related Pages in there. Note: in some browsers, the Javascript-based hierarchical list of documents in the sidebar at left will either take a long time to come up, or will cause Internet Explorer to ask you whether you want to abort a script that is taking to long (be sure to say that you don't want to abort). It takes only a few seconds to come up in Netscape 4.78 but can take close to a couple of minutes in other browsers.

Note: For academic and research use only! Most of this code is protected under a pending U.S. patent.

Anonymous SVN Access to the iLab Source Tree

The best mesthod is to use anonymous SVN access, as follows:

• svn checkout --username anonsvn svn://isvn.usc.edu/software/invt/trunk/saliency
• Password: itRocks
• svn update (while you are currently in the checked-out source tree) to later update and sync your copy to the master source tree to its latest version (this will update the contents of your current directory and its subdirectories).

If you are interested in contributing new pieces to the project, please let us know so that we can grant you full SVN access.

iLab C++ Neuromorphic Vision Toolkit VirtualBox Image

You can download a VirtualBox disk image with Mandriva 2010 installed, the source code for the toolkit, all additional Linux libraries, and everything pre-compiled and ready to go.

• VirtualBox Image with full source code: Mandriva-2010-pwp-i586-iLab.7z [1.1 GB]

Please see instructions on how to get started with VirtualBox and setting up a disk image here. The 2010 VirtualBox image here contains more than the 2009 one on that web page, namely:

• The entire source tree
• The whole source code has been compiled already, giving rise to a flury of executables
• Everything is in ~/saliency/ -- simply run svn update there to update to the latest version.
• Have a look at the various screenshots and documentation files here for tips on getting started.

iLab C++ Neuromorphic Vision Toolkit Source Tarballs

You can download daily snapshots of the source code here.

• Daily snapshot: saliency-current.tar.bz2